U.S. patent application number 12/576919 was filed with the patent office on 2011-04-14 for modeling distribution of emergency relief supplies for disaster response operations.
This patent application is currently assigned to INTERNATIONAL BUSINESS MACHINES CORPORATION. Invention is credited to Markus Ettl, Soumyadip Ghosh, Young M. Lee.
Application Number | 20110087514 12/576919 |
Document ID | / |
Family ID | 43855551 |
Filed Date | 2011-04-14 |
United States Patent
Application |
20110087514 |
Kind Code |
A1 |
Ettl; Markus ; et
al. |
April 14, 2011 |
MODELING DISTRIBUTION OF EMERGENCY RELIEF SUPPLIES FOR DISASTER
RESPONSE OPERATIONS
Abstract
A method and system to supply multiple items through a network
of inventory staging areas and distribution points by determining
inventory stocking levels for a staging area and distribution
points, and inventory shipments from said staging area to
distribution points and between the distribution points, using an
optimization formulation whose objective is to cover maximum
overall demand at the distribution points in a given period of
time, and to minimize total cost of meeting demand.
Inventors: |
Ettl; Markus; (Yorktown
Heights, NY) ; Ghosh; Soumyadip; (Yorktown Heights,
NY) ; Lee; Young M.; (Yorktown Heights, NY) |
Assignee: |
INTERNATIONAL BUSINESS MACHINES
CORPORATION
Armonk
NY
|
Family ID: |
43855551 |
Appl. No.: |
12/576919 |
Filed: |
October 9, 2009 |
Current U.S.
Class: |
705/7.25 ; 703/2;
705/29; 705/338 |
Current CPC
Class: |
G06Q 10/0875 20130101;
G06Q 10/08 20130101; G06Q 10/04 20130101; G06Q 10/06315 20130101;
G06Q 10/087 20130101; G06Q 10/08355 20130101 |
Class at
Publication: |
705/7.25 ;
705/29; 703/2; 705/338 |
International
Class: |
G06Q 10/00 20060101
G06Q010/00 |
Claims
1. A method for supplying multiple items through a network of
inventory staging areas and distribution points, comprising:
determining, using a processor, inventory stocking levels at a
staging area and distribution points, and inventory shipments
between said staging area and distribution points and between the
distribution points, using an optimization formulation whose
objective function is to cover maximum overall demand at the
distribution points in a given period of time, and to meet the said
demand at minimum total cost.
2. The method of claim 1, wherein the step of determining is
performed iteratively for each given period of time using data
associated with said each given period of time respectively as
input to the optimization formulation, at least one of the data
being dynamic data that changes over said each given period of
time.
3. The method of claim 1, wherein the objective to cover maximum
overall demand at the distribution points is formulated as
minimizing time for distributing total inventory through the
distribution points.
4. The method of claim 1, wherein the objective to cover maximum
overall demand at the distribution points is formulated as
minimizing a weighted average of individual supply depletion times
of the distribution points.
5. The method of claim 1, wherein the optimization formulation is
deterministic.
6. The method of claim 1, wherein the optimization formulation is
stochastic.
7. The method of claim 1, wherein the optimization formulation
includes: minimizing b .tau. + i { R LSA i t LSA i + j j .noteq. i
R CS ij t CS ij } , ##EQU00012## wherein b is shortage cost, .tau.
is time total inventory is depleted at the distribution points, i,
j represents i-th and j-th distribution points, R.sup.i.sub.LSA
represents amount of staging area inventory available for shipping
to distribution point i, t.sup.i.sub.LSA represents average time it
takes to ship from the staging area to the distribution point i,
R.sup.ij.sub.CS represents amount of shipment received at
distribution point j from distribution point i. t.sup.ij.sub.CS
represents average time it takes to ship from distribution point i
to the distribution point j.
8. The method of claim 1, wherein the optimization formulation
includes: minimizing b i E .tau. i + i { R LSA i t LSA i + j j
.noteq. i R CS ij t CS ij } , ##EQU00013## wherein b is shortage
cost, i E .tau. i ##EQU00014## is expected time value for total
inventory to be depleted at the distribution points, i, j
represents i-th and j-th distribution points, R.sup.i.sub.LSA
represents amount of staging area inventory available for shipping
to distribution point i, t.sup.i.sub.LSA represents average time it
takes to ship from the staging area to the distribution point i,
R.sup.ij.sub.CS represents amount of shipment received at
distribution point j from distribution point i, t.sup.ij.sub.CS
represents average time it takes to ship from distribution point i
to the distribution point j.
9. The method of claim 1, further including generating a schedule
for distribution using output from the optimization formulation,
generating a report of a schedule for distribution using output
from the optimization formulation, or combinations thereof.
10. A program storage device readable by a machine, tangibly
embodying a program of instructions executable by the machine to
perform a method of supplying multiple items through a network of
inventory staging areas and distribution points, comprising:
determining, by a processor, inventory stocking levels at a staging
area and distribution points, and inventory distribution from said
staging area and distribution points and between the distribution
points, using an optimization formulation whose objective is to
cover maximum overall demand at the distribution points in a given
period of time, and to minimize total cost of meeting the
demand.
11. The program storage device of claim 10, wherein the step of
determining is performed iteratively for each given period of time
using data associated with said each given period of time
respectively as input to the optimization formulation, at least one
of the data being dynamic data that changes over said each given
period of time.
12. The program storage device of claim 10, wherein the objective
to cover maximum overall demand at the distribution points is
formulated as minimizing time for distributing total inventory
through the distribution points.
13. The program storage device of claim 10, wherein the objective
to cover maximum overall demand at the distribution points is
formulated as minimizing a weighted average of individual supply
depletion times of the distribution points.
14. The program storage device of claim 10, wherein the
optimization formulation includes minimizing b .tau. + i { R LSA i
t LSA i + j j .noteq. i R CS ij t CS ij } , ##EQU00015## wherein b
is shortage cost, .tau. is time total inventory is depleted at the
distribution points, i, j represents i-th and j-th distribution
points, R.sup.i.sub.LSA represents amount of staging area inventory
available for shipping to distribution point i, t.sup.i.sub.LSA
represents average time it takes to ship from the staging area to
the distribution point i, R.sup.ij.sub.CS represents amount of
shipment received at distribution point j from distribution point
i. t.sup.ij.sub.CS represents average time it takes to ship from
distribution point i to the distribution point j.
15. The program storage device of claim 10, wherein the
optimization formulation includes minimizing b i E .tau. i + i { R
LSA i t LSA i + j j .noteq. i R CS ij t CS ij } , ##EQU00016##
wherein b is shortage cost, i E .tau. i ##EQU00017## is expected
time value for total inventory to be depleted at the distribution
points, i, j represents i-th and j-th distribution points,
R.sup.i.sub.LSA represents amount of staging area inventory
available for shipping to distribution point i, t.sup.i.sub.LSA
represents average time it takes to ship from the staging area to
the distribution point i, R.sup.ij.sub.CS represents amount of
shipment received at distribution point j from distribution point
i, t.sup.ij.sub.CS represents average time it takes to ship from
distribution point i to the distribution point j.
16. A system for supplying multiple items through a network of
inventory staging areas and distribution points, comprising: a
processor; an analytic engine operable execute on the processor and
to determine inventory stocking levels at a staging area and
distribution points, and inventory shipments from said staging area
and distribution points and between the distribution points, using
an optimization formulation whose objective is to cover maximum
overall demand at the distribution points in a given period of
time, and to minimize total cost of meeting demand.
17. The system of claim 16, wherein the analytic engine is operable
to determine the inventory stocking levels iteratively for each
given period of time using data associated with said each given
period of time respectively as input to the optimization
formulation, at least one of the data being dynamic data that
changes over said each given period of time.
18. The system of claim 16, wherein the objective to cover maximum
overall demand at the distribution points is formulated as
minimizing time for distributing total inventory through the
distribution points, or as minimizing a weighted average of
individual supply depletion times of the distribution points.
19. The system of claim 16, wherein the optimization formulation
includes minimizing b .tau. + i { R LSA i t LSA i + j j .noteq. i R
CS ij t CS ij } , ##EQU00018## wherein b is shortage cost, .tau. is
time total inventory is depleted at the distribution points, i, j
represents i-th and j-th distribution points, R.sup.i.sub.LSA
represents amount of staging area inventory available for shipping
to distribution point i, t.sup.i.sub.LSA represents average time it
takes to ship from the staging area to the distribution point i,
R.sup.ij.sub.CS represents amount of shipment received at
distribution point j from distribution point i. t.sup.ij.sub.CS
represents average time it takes to ship from distribution point i
to the distribution point j.
20. The system of claim 16, wherein the optimization formulation
includes minimizing b i E .tau. i + i { R LSA i t LSA i + j j
.noteq. i R CS ij t CS ij } , ##EQU00019## wherein b is shortage
cost, i E .tau. i ##EQU00020## is expected tune value for total
inventory to be depleted at the distribution points, i, j
represents i-th and j-th distribution points, R.sup.i.sub.LSA
represents amount of staging area inventory available for shipping
to distribution point i, t.sup.i.sub.LSA represents average time it
takes to ship from the staging area to the distribution point i,
R.sup.ij.sub.CS represents amount of shipment received at
distribution point j from distribution point i, t.sup.ij.sub.CS
represents average time it takes to ship from distribution point i
to the distribution point j.
Description
BACKGROUND
[0001] The present disclosure relates generally to distributing
supplies, and more particularly to modeling distribution of
emergency relief supplies for disaster operations. When disasters
occur (e.g., such as hurricane, earthquake, fire, bioterrorism, and
others), emergency supplies (e.g., water, meal, medicine,
generators, blankets, tarps, and others) need to be distributed to
victims on time. The distribution operations in those situations
(e.g., supply chain and dispensing) are unique because the
operations need to cover a large number of people (e.g., million of
victims) in a short period of time (e.g., a small number of hours
or days) under undesirable conditions for supply chain operation
(chaos, damaged and/or congested roadways, behavior of victims,
progression of disasters, many unknowns and uncertainties), and
serious consequences of an ineffective distribution plan (sickness,
social disorder, and others). It is usually a one time event of
short duration with limited opportunity for re-planning the supply
chain design.
[0002] Relief distribution supply chain operations differ from
typical industrial supply chains. Unlike standard stationary demand
distribution assumptions, relief operations need to take into
account a huge surge in demand within short notice. Unfavorable
logistical conditions for supply chain operations such as chaotic
traffic, damaged/congested roadways and chaotic behavior of demand
(victims) have to be explicitly considered. Additionally, lead time
requirements are short. Preparing for a large disaster such as
hurricane is difficult primarily because of the high uncertainty
involved in predicting where and when it will strike. Therefore,
operational research models to improve preparedness for and
response to major emergencies would be desirable.
[0003] A typical distribution for relief supplies starts from a
central warehousing or like, to a staging area from where the
supplies are distributed to individual point of distribution (POD)
locations. Persons needing the supplies collect them from the POD
locations. In most disaster situations, demand from victims
exhibits high uncertainty and variability. Disaster relief supplies
may reach different POD locations at different speed and with
different quantities, creating an imbalance between the supply and
demand. For example, certain PODs may experience shortage of
supplies and certain other POD location may experience surplus
during disaster response operations. Thus, it is desirable that
each POD location has the correct amount for distribution to the
demand at that location.
BRIEF SUMMARY OF THE INVENTION
[0004] A system and method for supplying multiple items through a
network of inventory staging areas and distribution points are
provided. The system, in one aspect, may include a processor and an
analytic engine that is operable and executable on the processor.
The analytic engine is operable to determine inventory stocking
levels at a staging area and distribution points, and inventory
shipments between said staging area and distribution points and
between the distribution points, using an optimization formulation
whose objective is to cover maximum overall demand at the
distribution points in a given period of time. The objective
function of the optimization formulation may also include
minimizing total cost of meeting the demand.
[0005] A method for supplying multiple items through a network of
inventory staging areas and distribution points, in one aspect, may
include determining inventory stocking levels at a staging area and
distribution points, and inventory shipments between said staging
area and distribution points and between the distribution points,
using an optimization formulation whose objective is to cover
maximum overall demand at the distribution points in a given period
of time. The objective function of the optimization formulation may
also include minimizing total cost of meeting the demand.
[0006] A program storage device readable by a machine, tangibly
embodying a program of instructions executable by the machine to
perform one or more methods described herein may be also
provided.
[0007] Further features as well as the structure and operation of
various embodiments are described in detail below with reference to
the accompanying drawings. In the drawings, like reference numbers
indicate identical or functionally similar elements.
BRIEF DESCRIPTION OF THE DRAWINGS
[0008] FIG. 1 is a block diagram illustrating a supply distribution
and cross shipping modeler of the present disclosure.
[0009] FIG. 2 illustrates a method of determining supply
distribution and cross shipping based in one embodiment of the
present disclosure.
[0010] FIG. 3A plots a sample path of a POD's usage of
inventory.
[0011] FIG. 3B plots a sample path of a POD's usage of inventory
over time under the stochastic demand model.
[0012] FIG. 4 illustrates an example of a computer system, in which
the systems and methodologies of the present disclosure may be
carried out or executed.
DETAILED DESCRIPTION
[0013] In one aspect, a Brownian Motion (BM) model is disclosed for
the relief supply demand to adequately capture the non-stationary,
volatility and correlation inherent in it. A tool is presented in
one embodiment of the present disclosure, which guides the relief
operations based on continually updating the operational plan by
solving a sequence of stochastic optimization problems. The model
may be used to effectively distribute supply coming into a staging
area in the disaster-hit area to local points of distribution
(POD), and also cross-level or cross-distribute among PODs, e.g.,
to attain the quick coverage of the affected populace and/or
areas.
[0014] In one aspect, an optimization model is defined that
attempts to find the best allocation of the incoming supply at the
staging area to the PODs that maximizes the objective of an agency
or the like responding to disaster situations. The optimization
model also allows for re-balancing of inventory between PODs
through cross-shipping. In one aspect, and unlike traditional
logistics of a typical supply chain, the optimization goal in the
model of the present disclosure focuses on the ability to quickly
cover the population demanding relief at the PODs.
[0015] FIG. 1 is a block diagram illustrating a supply distribution
and cross shipping modeler of the present disclosure. A relief
logistics optimizer 102 uses input such as demand forecast,
in-transit inventory, on-hand inventory, staging area incoming
supply, truck availability (or other transportation medium
availability), truck loading constraints (or other transportation
medium loading constraints), cross-shipping constraints, and road
travel conditions. The relief logistics optimizer 102 employs a
formulation to output relief supply distribution from the staging
area such as the local staging area (LSA) to different PODs, and
between different PODs. Example formulations of the optimizer 102
are explained further below. Logistics schedule calculator 104
converts the output to logistics schedule, which may include supply
distribution schedule from the staging area to different POD
locations 106 and cross-shipping schedule from a POD to another
POD. The logistics schedule calculator 104 may also generate a
report of the distribution, for example, that details expected
coverage, supply depletion time, and other performance
information.
[0016] FIG. 2 illustrates a method of determining supply
distribution and cross shipping in one embodiment of the present
disclosure. At 202, an optimization model (also referred to herein
interchangeable as an optimization engine or analytics engine), is
initialized. Initialization may include setting up a distance
matrix, POD parameters, truck (or other vehicle) availability and
loading constraints using input data such as the travel distance,
road conditions, vehicle availability, availability of one or more
loading docks, multiple-item shipping constraints, and others. The
data may be input by a user or read from available files as an
input to the model or combinations of both.
[0017] At 204, dynamic data, that is the information that changes
as the disaster happens and relief is demanded, is received. The
data may include, but is not limited to, demand characteristics,
i.e., number of persons needing or requesting the supplies at each
POD, on-hand and in-transit inventory at each POD and staging area,
and others. In one embodiment, the dynamic data collected at 204
applies to a periodic interval, e.g., next H hours. Thus, in one
embodiment, the decision that the model addresses may evolve over
time as the both the supply and demand may change over time. Supply
(or inventory) at a local staging area (LSA) is being brought in
from central warehouses located elsewhere outside the affected
zone. Demand evolves over time starting with the time of impact of
the disaster; typically, the demand starts very high at the
beginning of operations and tapers off.
[0018] At 206, an optimization model representation of the supply
delivery problem is constructed. The model can consider either a
deterministic or a stochastic model of demand evolution. In one
aspect, the decisions that the stochastic optimization formulation
address may include: 1) what should be the re-stocking level at
each POD, and 2) how should these re-stocking levels be attained.
The latter question may be answered by a combination of shipments
of supply either present or coming into the LSA that are diverted
to the PODs, and of cross-balancing shipments made between PODs to
help those that are facing imminent shortfall from those that have
a comfortable level of inventory. The decision process of the
present disclosure is simplified by formulating optimization
problems with a significantly shorter horizon (e.g., hours) than
the entire relief operations (days), and cyclically updating
decisions by re-solving with a pre-determined frequency. The
optimization problem in each cycle will determine the best set of
decisions that maximizes customer (demand) coverage over all PODs
given the total inventory available over H hours. The model and
(greedy) objective of each cycle is formulated in a manner that
obtains solutions close to global (in the time-scale sense)
optimality. The detailed algorithm for the optimization model is
explained further below.
[0019] At 208, optimization problem is solved using the constructed
model. At 210, the optimization solution is converted into
schedules and reports. For instance, the optimization model or
formulation may output inventory amounts to ship from LSA to one or
more PODs and from one POD to another POD, that satisfies the
constraints of the optimization model. Those output values may be
converted or formatted into actionable transportation schedules
that implement the determined shipment, and reports that summarize
the current and expected performance of the supply delivery system,
identify potential supply shortcomings, etc.
[0020] At 212, if the relief distribution operation has ended, the
analytic engine stops, otherwise, the steps 204-210 repeat for the
next interval period.
[0021] The optimization formulation is now explained in more
detail. In the following description, trucks are referred to as the
medium for transporting supplies as example. It should be
understood, however, that the present disclosure does not limit
transportation vehicles to only trucks; rather other medium may be
utilized for transporting the supplies. The cycle-horizon H is
fixed to be of the order of the average time it takes for a truck
(or another transportation medium) to make a round-trip between LSA
and the PODs. Thus, the limitation on the number of available
trucks becomes a natural constraint on the maximum shipments
allowed in each horizon. Standard inventory theory dynamics are
assumed for this discretized model: each POD starts with a specific
inventory-at-hand I.sup.i and a customer-queue Q.sup.i, any
shipments due at the POD arrives at the beginning, the demand for
the timeslot is realized over time, and service is provided
continuously throughout the slot. The constrained service delivery
rate is a restriction faced by relief operations in practice and is
explicitly addressed in the model. We model this using a maximum
service rate S.sup.i. Service is conservative, i.e., continues at
maximum rate without delay as long as any inventory is available. A
finite limited service rate implies that both inventory-at-hand
I.sup.i and customer-queue (back-orders) Q.sup.i can be non-zero,
and also lets us relate the inventory and queue lengths over the
horizon given the inventory distribution decisions. The optimal
decision formulation then determines on the best allocation of the
total inventory available to the system amongst the PODs. Call
I.sup.LSA the total available (current and expected over H hours)
inventory at the LSA. In addition the inventories I.sup.i at the
POD i can also be re-distributed. Let R.sup.i represent the change
in inventory at POD i as a result of our distribution decisions.
Note that R.sup.i can be negative, representing cross-shipments of
inventory from POD i. For the ease of modeling the servicing
dynamics of each POD, we assume that this re-distribution is
instantaneous. This is reasonable given the limit on the throughput
S.sup.i<.infin., and the anticipation that in most instances of
this problem I.sup.i>0 to start with. The optimization objective
in one embodiment, however, may use the length of delivery to
penalize cross-shipments across large distances.
[0022] The logistical constraints that need to be placed on the
distribution variable R' are now described.
Logistical Constraints
[0023] Total received replenishment R.sup.i comes from either LSA
or other PODs via redistributions.
[0024] Define a set of variables R.sup.ij.sub.CS for each
i.apprxeq.j, each representing the total cross-shipment received at
POD j from POD i. Let R.sup.i.sub.LSA represent the amount of LSA's
inventory available currently or over the next H hours that is
shipped to POD i. Balancing R.sup.i with the R.sup.ij.sub.CS and
R.sup.i.sub.LSA, we have
.A-inverted. i , R i = R LSA i + j j .noteq. i ( R CS ji - R CS ij
) ( 1 ) ##EQU00001##
[0025] The optimal solution is penalized to ensure it picks only
one of R.sup.ij.sub.CS or R.sup.ij.sub.CS to be non-zero.
Additional trucking constraints may apply to the LSA and POD to POD
cross-shipments. A typical truckload for LSA shipment may be a full
18-wheeler with a total capacity of TC.sub.LSA KiloLitres, where
typically TC.sub.LSA=16. Further, it may be that only full
truckloads are mobilized in order to minimize the number of large
trucks on the already-fragile road infrastructure. Define N(H) to
be the total full truckloads of inventory (supply) available at the
LSA over the next H hours. Define for each POD i an integer-valued
variable x.sup.i.sub.LSA representing the number of truckloads sent
from LSA to POD i. These variables satisfy
i x LSA i .ltoreq. N ( H ) R LSA i = x LSA i TC , .A-inverted. i x
LSA i .di-elect cons. { 0 , 1 , 2 , , N ( H ) } .A-inverted. i . (
2 ) ##EQU00002##
[0026] The N(H) itself accounts for the limited trucks available in
the next H hours for LSA logistics operations.
[0027] For the cross-shipment part, both the truck-size and the
number of trucks available may constrain the total cross shipments.
Define NT.sub.CS to be the total cross-shipping trucks available.
These are typically smaller than 18-wheelers, and their carrying
capacity TCCS is a fraction of the LSA load TC.sub.LSA.
Additionally, POD inventories may be held in the pallets they came
in from the LSA, and thus cross-shipment loads may be multiples of
a standard pallet size P.sub.CS. For each POD i define
integer-valued variables y.sup.ij.sub.CS and x.sup.ij.sub.CS where
the first represents the number of pallets cross-shipped from POD i
to j, and the second represents the number of trucks needed to
carry this load. The following constraints round out the
cross-shipment model:
R CS ij = y CS ij P CS .A-inverted. i , j , i .noteq. j y CS ij P
CS .ltoreq. x CS ij TC CS .A-inverted. i , j , i .noteq. j i , j i
.noteq. j x CS ij .ltoreq. NT CS y CS ij .di-elect cons. { 0 , 1 ,
2 , , TC CS / P CS NT CS } .A-inverted. i , j , i .noteq. j y CS ij
.di-elect cons. { 0 , 1 , 2 , , NT CS } .A-inverted. i , j , i
.noteq. j . ( 3 ) ##EQU00003##
[0028] The following now describes two objects in one embodiment
that are minimized in a balanced manner in the optimization
formulations.
OBJECTIVES
[0029] In one embodiment, the goal of the formulation is to
maximize the overall coverage achieved using the supply available
through the next H hours. Define for POD i the quantity .tau..sup.i
to be the time to drain out I.sup.i+R.sup.i amount of inventory
starting from the current time. The drain-out time .tau..sup.i
depends on 1) the amount I.sup.i+R.sup.i, 2) the servicing rate
S.sup.i, and 3) the demand process serviced by the POD. Call Di(t)
POD i's cumulative demand by t, starting with Q.sup.i at current
time t=0. Then, a good proxy for our customer-coverage maximization
goal using the current inventory-at-hand may be achieved by
minimizing a function of the individual .tau..sup.i's. The
following two objects may be minimized:
.tau. = max .tau. i , or ( 4 ) .tau. = i w i .tau. i . ( 5 )
##EQU00004##
[0030] With the first definition (4), .tau. stands for the time
when the last of the total inventory (I.sup.LSA+.tau..sub.iI.sup.i)
is used up. The second objective (5) minimizes a weighed average of
the individual supply depletion times .tau..sup.i. This scheme of
solving optimization problems sequentially can then be interpreted
as one which tries to find the quickest way to dissipate short-term
available supply through the given POD network.
[0031] Another objective minimizes the total cost of shipment,
i { R LSA i t LSA i + j j .noteq. i R CS ij t CS ij } , ( 6 )
##EQU00005##
where t.sup.i.sub.LSA is the average time it takes to truck from
the LSA to POD i, and t.sup.ij.sub.CS is the average shipping time
from POD i to j. The triangle-inequality is assumed to hold for any
three locations, but note that t.sup.ij.sub.CS and t.sup.ij.sub.CS
may not coincide because of various local restrictions (one-ways,
etc.) imposed on the degraded road infrastructure. These penalties
ensure that only necessary shipments are considered and frivolous
shipments between far-flung PODs, and cross-shipments between each
pair i, j in both directions are avoided.
[0032] The two objectives may be jointly minimized, each weighed
appropriately. We associate an inventory shortage cost b with the
coverage objective.
Deterministic Model
[0033] FIG. 3A plots a sample path of a POD's usage of inventory.
At time t=0 (current time), a backorder of Q.sup.i exists, and the
total distribution to POD i is (I.sup.i+R.sup.i). The POD starts
consuming inventory at the linear rate S.sup.i. On the other hand,
demand grows at the linear rate .lamda..sup.i, and thus the total
demand faced by POD i at time t is D.sup.i(t)=Q.sup.i+.lamda..sup.i
t. Let f.sup.i (t):=min{S.sup.it,D.sup.i(t)} be the drain-out
process over time. Then, the earliest drain-out time .tau..sup.i of
this POD i starting with the current distribution plan is
.tau..sup.i:=min{t|f.sup.i(t)=I.sup.i+R.sup.i}.
[0034] Since both the service and demand functions are
non-decreasing (.lamda..sup.i.gtoreq.0), so is f.sup.i and then we
have the simpler .tau..sup.i definition
f(.tau..sup.i)=I.sup.i+R.sup.i. The function f is piecewise linear,
and so the constraint can be reformulated as a set of linear
inequalities in terms of the variables .tau..sup.i and R.sup.i.
Define v.sup.i.sub.S and v.sup.i.sub.D to be the intersection times
of the service and demand curves with the inventory level
I.sup.i+R.sup.i respectively, as illustrated in FIG. 3A. The time
v.sup.i.sub.D=+.infin. if the demand accumulation curve does not
intersect with the inventory level, and can also be negative. The
earliest drain-out time .tau..sup.i=max {0, v.sup.i.sub.S,
v.sup.i.sub.D}.
[0035] With the objective (4) of minimizing the maximum of the POD
drain-out times, a linear programming then allows us to use the
following set of equivalent constraints in place of the piece-wise
definition of .tau..sup.i:
.tau..sup.i.gtoreq.0
.gtoreq.v.sup.i.sub.D=(I.sup.i+R.sup.i-Q.sup.i)/.lamda..sup.i
.gtoreq.v.sup.i.sub.S=(I.sup.i+R.sup.i)/S.sup.i. (7)
[0036] For any optimal solution {.tau..sup.i.sub.*, R.sup.i.sub.*,
.A-inverted.i} to the first objective (4), the POD that attains the
maximum i*=arg max.sub.i .tau..sup.i.sub.* will satisfy
.tau..sup.i*.sub.*=max {0, v.sup.i*.sub.D, v.sup.i*.sub.S}. If the
.tau..sup.i.sub.* for any POD i.noteq.i* does not match the
corresponding maxima {0, v.sup.i.sub.S, V.sup.i.sub.D}, a modified
solution can be obtained by setting those value to equal the
corresponding maximum value without affecting the solution
cost.
[0037] When the primary objective is of form (5), a Special Ordered
Set to model .tau..sup.i may be used.
[0038] In one embodiment, the complete optimization formulation for
deterministic demand is:
min b .tau. + i { R LSA i t LSA i + j j .noteq. i R CS ij t CS ij }
. ##EQU00006##
[0039] such that [0040] I.sup.i=<data>, Q.sup.i=<data>
(initial conditions) [0041] supply distribution R.sup.i to i as
defined by (1), (2) and (3) [0042] measure .tau. as defined in (4)
or (5) constraints on individual .tau..sup.i of form (7).
[0043] In the above optimization formulation, represents customer
satisfaction criteria, i.e., one or more decision objectives,
and
i { R LSA i t LSA i + j j .noteq. i R CS ij t CS ij }
##EQU00007##
represents the transportation cost from (6).
Stochastic Model for Relief Supply Demand
[0044] The demand faced by a relief delivery operation is very
different from the standard supply chain models. Demand is highly
non-stationary, time-dependent and volatile. We use a Brownian
Motion model of the demand faced by the PODs. The cumulative demand
process D(t) is
a I-dimensional process that is defined in terms of its increment
at time t
dD(t)=.lamda.(t)+A(t)dB(0,tI)
where .lamda.(.cndot.), a I-vector, represents the time-dependent
rate of change of demand, A(t) is the I.times.I standard deviation
process, and B(0, I) is a zero-mean, uncorrelated Brownian Motion
with standard deviations of one. The mean .lamda. (.cndot.) changes
with the time-of-day. One can expect it to be high during the AM
hours as disaster affected families look to collect their supplies
for the day, and dies down slowly over the day. This pattern
broadly lasts for the duration of the relief-operations, and falls
off over time as more of the affected civic infrastructure comes
back up. Moreover, the affected region can be classified into bands
where the effect is heavier or lighter depending on the severity of
damage to the region. The standard deviation process A(t) helps
capture the significant variation that might occur, for instance
due to queue abandonment in favor of another nearby POD. We can
also capture the significant correlation that might exist between
the demand faced by close-by PODs: for example, demand fielded by
adjacent PODs can be correlated because the local populace might
visit multiple nearby PODs looking to obtain the quickest
service.
[0045] Our cyclic decomposition approach to the decision-making
process allows us to simplify the demand model slightly and assume
that the mean and standard-deviation are constant over the next H
hours from the current time. Thus, the demand growth faced by the
stochastic optimization problem may be of the form:
D(t)=D(0)+.lamda.t+AB(0,tI),
where time t=0 represents the current time, and D(0) the current
demand (e.g., customer-queue). We let Q represent D(0). Model with
Stochastic Demand
[0046] FIG. 3B plots a sample path of a POD's usage of inventory
over time under the stochastic demand model. The demand in this
case grows stochastically as Q+D(t) where the i-th component is
D.sup.i(t)=Q.sup.i+.lamda..sup.it+A.sub.iB(0,tI),
and A.sub.i is the i-th row of the matrix A. Let
f.sup.i(t)min{S.sup.it,D.sup.i (t)} represent the stochastic
drain-out process. The f.sup.i (t) process is a scaled and rotated
standard Reflected Brownian Motion (RBM); the rotation is set by
the service rate S.sup.i (t) and the correlation matrix A. Then,
the earliest drain-out time .tau..sup.i of this POD i starting with
the current distribution plan is the first hitting time of the
stochastic RBM process f.sup.i:
.tau..sup.imin{t|f.sup.i(t)=I.sup.i+R.sup.i}.
[0047] This formulation for optimization under stochastic demand
may allow us to expand the expectation of the .tau..sup.i
explicitly so that the final formulation solved is deterministic.
This may avoid the computational penalty incurred by any solution
procedure that handles stochastic formulations via scenario
generation or stochastic approximations. In one embodiment, we use
an approximation to E.tau..sup.i, the expected value of
.tau..sup.i. Analogous to the deterministic case described above,
define:
v.sup.i.sub.Smin{t:S.sup.i.sub.t=I.sup.i+R.sup.i}
v.sup.i.sub.Dmin{t:I.sup.i+R.sup.i=Q.sup.i+.lamda..sup.it+AB(0,tI)}
v.sup.i.sub.DSmin{t:S.sup.it=Q.sup.i+.lamda..sup.it+AB(0,tI)}.
[0048] The term v.sup.i.sub.S is deterministic as in the earlier
case, and has value (I.sup.i+R.sup.i)/S.sup.i. The quantities
v.sup.i.sub.D and v.sup.i.sub.DS are stochastic and by definition
are stopping times associated with the demand growth
Q.sup.i+D.sup.i(t). An approximation for the earliest draining time
.tau..sup.i is given by
.tau. i = { v D i if v DS i .ltoreq. v S i , v S i otherwise ( 8 )
##EQU00008##
and the expected value of .tau..sup.i can be calculated as
E.tau..sup.i=Ev.sup.i.sub.DP(v.sup.i.sub.DS.ltoreq.v.sup.i.sub.S)+v.sup.-
i.sub.S(1-P(v.sup.i.sub.DS.ltoreq.v.sup.i.sub.S)). (9)
[0049] The expected value Ev.sup.i.sub.D is obtained by standard
techniques for Brownian Motion (BM) first hitting times. The
probability P(v.sup.i.sub.DS.ltoreq.v.sup.i.sub.S) can also be
calculated based on earliest exit times of Brownian Motions from
wedges. V. Fabian in "Note on Anderson's sequential procedures with
triangular boundary" (The Annals of Statistics, 2:170-176, 1974),
provides an exact expression for the probability that a BM with a
known drift starting from the origin exits out of a triangular
region symmetrically drawn around the time-axis (i.e., the BM value
0) via the expected arm, which is the arm that lies in the
direction of the drift. This result is used extensively in the
ranking-and-selection approach to discrete stochastic optimization
to bound the probability of correctly rejecting inferior points
from the finite parameter set (see, S.-H. Kim and B. L. Nelson. "On
the asymptotic validity of fully sequential selection procedures
for steady state simulation", Operations Research, 54:475-488,
2006). Note that the expression in the equations (9) can be
non-linear in the distribution variables R.sup.i.
[0050] We seek to minimize a function .tau. of these .tau..sup.i
over i.epsilon.I. This is their maximum if (4) is followed. The
expectation of this maximum cannot however be written down readily
in closed form. In another embodiment, primary objective to
minimize may be chosen to be E.tau.=.SIGMA..sup.i.tau..sup.i as in
(5). This objective readily yields a closed form expression for all
terms involved and thus provides a deterministic formulation to
solve. The complete formulation with this objective would be:
min b i E .tau. i + i { R LSA i t LSA i + j j .noteq. i R CS ij t
CS ij } . ##EQU00009##
[0051] such that [0052] I.sup.i=<data>, Q.sup.i=<data>
(initial conditions) [0053] supply distribution R.sup.i to i as
defined by (1), (2) and (3) [0054] individual E.tau..sup.i is
defined by equation of form (9).
[0055] In the above optimization formulation
i E .tau. i ##EQU00010##
represents customer satisfaction criteria, i.e., one or more
decision objectives, and
i { R LSA i t LSA i + j j .noteq. i R CS ij t CS ij }
##EQU00011##
represents the transportation cost from (6).
[0056] The solution to these optimization problems determines the
target set of inventory levels at each POD and a plan to
redistribute inventory from the staging area and the PODs, and
between PODs. This information may be then converted into a
dynamically updating actionable transportation schedule where
vehicles are assigned specific tasks in implementing the
re-distribution targets. Schedule creation takes into consideration
factors like location- and time-availability of free transportation
vehicles, driver availability, and any logistical constraints on
the local manpower (minimum rest period for drivers, and other
factors) and facilities.
[0057] The above algorithms are described as examples only and thus
the invention is not limited to using only that algorithm. Other
algorithm using one or more or different combination of dynamic
parameters such as those described above, may be utilized to
provide cross shipping decisions.
[0058] As will be appreciated by one skilled in the art, aspects of
the present invention may be embodied as a system, method or
computer program product. Accordingly, aspects of the present
invention may take the form of an entirely hardware embodiment, an
entirely software embodiment (including firmware, resident
software, micro-code, etc.) or an embodiment combining software and
hardware aspects that may all generally be referred to herein as a
"circuit," "module" or "system." Furthermore, aspects of the
present invention may take the form of a computer program product
embodied in one or more computer readable medium(s) having computer
readable program code embodied thereon.
[0059] Any combination of one or more computer readable medium(s)
may be utilized. The computer readable medium may be a computer
readable signal medium or a computer readable storage medium. A
computer readable storage medium may be, for example, but not
limited to, an electronic, magnetic, optical, electromagnetic,
infrared, or semiconductor system, apparatus, or device, or any
suitable combination of the foregoing. More specific examples (a
non-exhaustive list) of the computer readable storage medium would
include the following: an electrical connection having one or more
wires, a portable computer diskette, a hard disk, a random access
memory (RAM), a read-only memory (ROM), an erasable programmable
read-only memory (EPROM or Flash memory), an optical fiber, a
portable compact disc read-only memory (CD-ROM), an optical storage
device, a magnetic storage device, or any suitable combination of
the foregoing. In the context of this document, a computer readable
storage medium may be any tangible medium that can contain, or
store a program for use by or in connection with an instruction
execution system, apparatus, or device.
[0060] A computer readable signal medium may include a propagated
data signal with computer readable program code embodied therein,
for example, in baseband or as part of a carrier wave. Such a
propagated signal may take any of a variety of forms, including,
but not limited to, electro-magnetic, optical, or any suitable
combination thereof. A computer readable signal medium may be any
computer readable medium that is not a computer readable storage
medium and that can communicate, propagate, or transport a program
for use by or in connection with an instruction execution system,
apparatus, or device.
[0061] Program code embodied on a computer readable medium may be
transmitted using any appropriate medium, including but not limited
to wireless, wireline, optical fiber cable, RF, etc., or any
suitable combination of the foregoing.
[0062] Computer program code for carrying out operations for
aspects of the present invention may be written in any combination
of one or more programming languages, including an object oriented
programming language such as Java, Smalltalk, C++ or the like and
conventional procedural programming languages, such as the "C"
programming language or similar programming languages. The program
code may execute entirely on the user's computer, partly on the
user's computer, as a stand-alone software package, partly on the
user's computer and partly on a remote computer or entirely on the
remote computer or server. In the latter scenario, the remote
computer may be connected to the user's computer through any type
of network, including a local area network (LAN) or a wide area
network (WAN), or the connection may be made to an external
computer (for example, through the Internet using an Internet
Service Provider).
[0063] Aspects of the present invention are described below with
reference to flowchart illustrations and/or block diagrams of
methods, apparatus (systems) and computer program products
according to embodiments of the invention, It will be understood
that each block of the flowchart illustrations and/or block
diagrams, and combinations of blocks in the flowchart illustrations
and/or block diagrams, can be implemented by computer program
instructions. These computer program instructions may be provided
to a processor of a general purpose computer, special purpose
computer, or other programmable data processing apparatus to
produce a machine, such that the instructions, which execute via
the processor of the computer or other programmable data processing
apparatus, create means for implementing the functions/acts
specified in the flowchart and/or block diagram block or
blocks.
[0064] These computer program instructions may also be stored in a
computer readable medium that can direct a computer, other
programmable data processing apparatus, or other devices to
function in a particular manner, such that the instructions stored
in the computer readable medium produce an article of manufacture
including instructions which implement the function/act specified
in the flowchart and/or block diagram block or blocks.
[0065] The computer program instructions may also be loaded onto a
computer, other programmable data processing apparatus, or other
devices to cause a series of operational steps to be performed on
the computer, other programmable apparatus or other devices to
produce a computer implemented process such that the instructions
which execute on the computer or other programmable apparatus
provide processes for implementing the functions/acts specified in
the flowchart and/or block diagram block or blocks.
[0066] The flowchart and block diagrams in the figures illustrate
the architecture, functionality, and operation of possible
implementations of systems, methods and computer program products
according to various embodiments of the present invention. In this
regard, each block in the flowchart or block diagrams may represent
a module, segment, or portion of code, which comprises one or more
executable instructions for implementing the specified logical
function(s). It should also be noted that, in some alternative
implementations, the functions noted in the block may occur out of
the order noted in the figures. For example, two blocks shown in
succession may, in fact, be executed substantially concurrently, or
the blocks may sometimes be executed in the reverse order,
depending upon the functionality involved. It will also be noted
that each block of the block diagrams and/or flowchart
illustration, and combinations of blocks in the block diagrams
and/or flowchart illustration, can be implemented by special
purpose hardware-based systems that perform the specified functions
or acts, or combinations of special purpose hardware and computer
instructions.
[0067] Referring now to FIG. 4, the systems and methodologies of
the present disclosure may be carried out or executed in a computer
system that includes a processing unit 2, which houses one or more
processors and/or cores, memory and other systems components (not
shown expressly in the drawing) that implement a computer
processing system, or computer that may execute a computer program
product. The computer program product may comprise media, for
example a hard disk, a compact storage medium such as a compact
disc, or other storage devices, which may be read by the processing
unit 2 by any techniques known or will be known to the skilled
artisan for providing the computer program product to the
processing system for execution.
[0068] The computer program product may comprise all the respective
features enabling the implementation of the methodology described
herein, and which--when loaded in a computer system--is able to
carry out the methods. Computer program, software program, program,
or software, in the present context means any expression, in any
language, code or notation, of a set of instructions intended to
cause a system having an information processing capability to
perform a particular function either directly or after either or
both of the following: (a) conversion to another language, code or
notation; and/or (b) reproduction in a different material form.
[0069] The computer processing system that carries out the system
and method of the present disclosure may also include a display
device such as a monitor or display screen 4 for presenting output
displays and providing a display through which the user may input
data and interact with the processing system, for instance, in
cooperation with input devices such as the keyboard 306 and mouse
device 8 or pointing device. The computer processing system may be
also connected or coupled to one or more peripheral devices such as
the printer 10, scanner (not shown), speaker, and any other
devices, directly or via remote connections. The computer
processing system may be connected or coupled to one or more other
processing systems such as a server 10, other remote computer
processing system 14, network storage devices 12, via any one or
more of a local Ethernet, WAN connection, Internet, etc. or via any
other networking methodologies that connect different computing
systems and allow them to communicate with one another. The various
functionalities and modules of the systems and methods of the
present disclosure may be implemented or carried out distributedly
on different processing systems (e.g., 2, 14, 16), or on any single
platform, for instance, accessing data stored locally or
distributedly on the network.
[0070] The terminology used herein is for the purpose of describing
particular embodiments only and is not intended to be limiting of
the invention. As used herein, the singular forms "a", "an" and
"the" are intended to include the plural forms as well, unless the
context clearly indicates otherwise. It will be further understood
that the terms "comprises" and/or "comprising," when used in this
specification, specify the presence of stated features, integers,
steps, operations, elements, and/or components, but do not preclude
the presence or addition of one or more other features, integers,
steps, operations, elements, components, and/or groups thereof.
[0071] The corresponding structures, materials, acts, and
equivalents of all means or step plus function elements, if any, in
the claims below are intended to include any structure, material,
or act for performing the function in combination with other
claimed elements as specifically claimed. The description of the
present invention has been presented for purposes of illustration
and description, but is not intended to be exhaustive or limited to
the invention in the form disclosed. Many modifications and
variations will be apparent to those of ordinary skill in the art
without departing from the scope and spirit of the invention. The
embodiment was chosen and described in order to best explain the
principles of the invention and the practical application, and to
enable others of ordinary skill in the art to understand the
invention for various embodiments with various modifications as are
suited to the particular use contemplated.
[0072] Various aspects of the present disclosure may be embodied as
a program, software, or computer instructions embodied in a
computer or machine usable or readable medium, which causes the
computer or machine to perform the steps of the method when
executed on the computer, processor, and/or machine. A program
storage device readable by a machine, tangibly embodying a program
of instructions executable by the machine to perform various
functionalities and methods described in the present disclosure is
also provided.
[0073] The system and method of the present disclosure may be
implemented and run on a general-purpose computer or
special-purpose computer system. The computer system may be any
type of known or will be known systems and may typically include a
processor, memory device, a storage device, input/output devices,
internal buses, and/or a communications interface for communicating
with other computer systems in conjunction with communication
hardware and software, etc.
[0074] The terms "computer system" and "computer network" as may be
used in the present application may include a variety of
combinations of fixed and/or portable computer hardware, software,
peripherals, and storage devices. The computer system may include a
plurality of individual components that are networked or otherwise
linked to perform collaboratively, or may include one or more
stand-alone components. The hardware and software components of the
computer system of the present application may include and may be
included within fixed and portable devices such as desktop, laptop,
server. A module may be a component of a device, software, program,
or system that implements some "functionality", which can be
embodied as software, hardware, firmware, electronic circuitry, or
etc.
[0075] The embodiments described above are illustrative examples
and it should not be construed that the present invention is
limited to these particular embodiments. Thus, various changes and
modifications may be effected by one skilled in the art without
departing from the spirit or scope of the invention as defined in
the appended claims.
* * * * *